Machine Learning2021ML Engineer
Bank Churn Prediction
PEP8, tested, logged production-style classifier
Churn model on bank_data.csv with modular EDA. Production-ready structure: fully tested, logged, and PEP8-compliant (Pylint / Autopep8).
The Problem & Engineering Constraint
The Core Challenge
Notebook-style churn analysis is hard to deploy. The code needed tests, logs, and style gates before it could be treated as a service.
Technical Architecture & Approach
Engineering Solution & Implementation
Modular Python with Conda, pandas/numpy/scipy EDA, scikit-learn models, Matplotlib/Seaborn plots, a logger, Pylint, and Autopep8.
Measured Production Impact
Verified Outcomes & Deliverables
PEP8-compliant, logged, and tested churn pipeline.
Modular EDA ready for later deployment.
Technologies & Components
System Tooling & Technologies
PythonScikit-LearnPandasNumPyPylintAutopep8